Clustering method and device for support vector machine

A support vector machine and clustering method technology, applied in the field of clustering methods and devices, can solve problems such as discrete and discontinuous data points of scatter plots, and achieve the effect of easy transplantation and strong adaptability

Active Publication Date: 2010-12-22
SHENZHEN MINDRAY BIO MEDICAL ELECTRONICS CO LTD +1
View PDF2 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Sometimes there are many data points in a sample, and the above density-based method may work, but many times the data points on the scatter plot are not continuous and very discrete

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Clustering method and device for support vector machine
  • Clustering method and device for support vector machine
  • Clustering method and device for support vector machine

Examples

Experimental program
Comparison scheme
Effect test

example 1

[0028] Example 1: Divider (identification of abnormal blood samples)

[0029] Clinically, patients with abnormal blood cells are usually screened out in routine blood examinations, and then blood and bone marrow smears are confirmed by microscopy. In blood cell analyzers based on flow cytometry, the scatter diagram of the test results usually has two configurations: using a combination of forward scattered light FSC and side scattered light SSC, or fluorescent FL and side scattered light The combination of SSC. The two-dimensional scatter plot distribution of blood samples from leukemia patients is significantly different from that of normal people. The scatter plot can be used to screen out such scatter plots with abnormal distribution patterns for the doctor to further comprehensively judge.

[0030] In this embodiment, a scattergram combining fluorescent FL and side-scattered light SSC is taken as an example to illustrate a two-class classification method for the shape recogni...

example 2

[0055] Example 2: Two-dimensional multi-classification (classification of routine blood white blood cells)

[0056] In clinical blood tests, white blood cells can be further divided into neutrophils, eosinophils, basophils, lymphocytes, and monocytes. Figure 5 It is a scatter chart of the blood cell test results of a normal person. It is clinically necessary to automatically classify and count the five types of white blood cells in the scatter chart.

[0057] Image 6 Described figure 1 The second step shown is the main process of the training step. By selecting training blood samples and manually adding training points, the support vector machine classification model is trained. The main process is as follows:

[0058] In step S600, browse the blood sample library and select a blood sample suitable for training the model. In step S602, it is judged whether a selected sample is used as a training sample. The principle followed can refer to the principle of selecting the blood sampl...

example 3

[0086] Example 3: Three-dimensional or high-dimensional classification

[0087] In order to further obtain various cell subgroups in blood cells (for example, lymphocytes can be further subdivided into T lymphocytes and B lymphocyte subgroups), more detection information needs to be obtained. Multicolor fluorescence analysis based on flow cytometry is usually used for analysis. At this time, multiple dimensions of detection information will be obtained.

[0088] Due to the characteristics of the support vector machine classifier, the method based on principal component analysis and support vector machine of this embodiment is more suitable for such high-dimensional signal clustering analysis.

[0089] Such as Figure 8 As shown, when there are three or more attributes that can be obtained for each blood cell that represent different characteristics of the cell, there are many ways to construct the input attributes of the support vector machine. Since support vector machines can ove...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses a cell clustering method and a device. The method comprises a generating step of generating a transformation matrix by using a collected blood sample, a training step of obtaining a support vector machine classified model of a support vector machine through training by using the chosen training blood sample, and a classifying step of transforming a scatter diagram formed by roughing an original data with the transformation matrix, obtaining a feature attribute vector of the cell by adding the data of a cell channel, and inputting the feature attribute vector to the trained support vector machine model so as to obtain a classifying result of the cell. The method and device of the embodiment of the invention have strong adaptability, are capable of classifying various types of samples and can be easily transplanted and applied to other related classifications.

Description

Technical field [0001] The invention relates to a clustering method and device, in particular to a clustering method and device of a support vector machine. Background technique [0002] Flow Cytometry (FCM) is a technology that performs multi-parameter rapid quantitative analysis and sorting of cells or other biological particles in liquid flow. In flow cytometry measurement, two scattering directions of scattered light are commonly used for measurement, namely forward angle scattering (FSC) light and side scattering (SSC) light. [0003] In some existing technologies, the "gating" method is used to identify, classify and count cells in blood samples. This method is relatively easy to implement, but one of the fatal weaknesses of cell classification using pre-defined boundaries is that when the morphology of the blood sample changes, the recognition will be wrong. [0004] Other existing technologies have made improvements to the above-mentioned shortcomings, such as the use of a ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G01N21/49G01N21/64
Inventor 吴鹏赵军
Owner SHENZHEN MINDRAY BIO MEDICAL ELECTRONICS CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap